Convergence of jet flow simulation
 

Hello!
I am simulating for a jet flow, that is two phases; air and water. The complete simulation consists of a open channel flow with high head, the water in the channel flows out in the air to land in a water basin. The water flowing out of the channel forms a jet flow. To decrease the number of cells in the model it was divided in two parts; one for the channel and one for the jet flow out from the channel. There is no problem with the channel simulation, and when that simulation was converged I wrote a profile-file containing all the information that I needed to run the jet flow simulation.

Now, the problem is that it takes very long time to run the jet flow simulation. From start, the cells of the entire simulation area (except from some boundaries) are containing around 50 % water. Pretty soon the jet flow gets the shape that it should have.

I run an unsteady simulation with 0.05 time step and 40 iterations per time step. What I wonder is in which order Fluent (I use Fluent 6.2) solves the NS equations. Because it doesn't seem like all cells converge at the same time step. For example, I monitor the volume fraction at different points in x-direction and quite fast it has converged at around 300 m in x-direction. However at around 500 m it keeps changing for many time steps, over 2000, after it has stabilized at 300 m. This means that the simulation will take about a week to converge (at 500 m). The model has around 180'000 cells.

Can anyone help explain to me why this is happening?

Thankful for help!